I correct with APL only the full range.
The speakers are already time-aligned using REW and loopback measurements, level matched and i am using linear-phase crossovers. I generated them in Rephase with some minor EQ to correct huge dips of nulls and then, on top of that, i measure the aforementioned cube area with APL. I get about 80-100 measurements per side on APL.
By the way, level matching in REW doesn't may not necessarily work in how APL "sees" those level. Before applying the correction, i may re-do measurements in APL until levels in the power response across all the speakers in a three way look more or less even.
Stabme, when you do the Workshop measurements for level matching of the power responses of all drivers , do you use the area suggested by Raimonds or your cube?
In Workshop what accuracy of level matching you aim at between L/R channel and across the three ways, how many dBs ? I have found it quite difficult to match all six levels of acoustic power frequency response for a 3way system.